Patients rank toxicity against progression free survival in second-line treatment of advanced renal cell carcinoma

Figures & data

Table 1.  Attributes and levels with descriptions for the choice questions.

Attribute	Attribute description	Levels	Level description
How long the medicine will keep the cancer from getting worse (progression-free survival)	An important goal of cancer medicines is to keep your cancer from getting worse. Sometimes cancer medicines will also improve cancer symptoms, but sometimes they may only keep symptoms from getting worse. Even if the medicine keeps the cancer from getting worse, it may stop working at some point and the cancer will get worse.	10 months	Not applicable
		5 months
		3 months
Feeling weak or tired (fatigue)	When you have cancer you may experience some weakness or tiredness (you have less energy or strength). Some cancer medicines may make you feel even more weak or tired. If you are weak or tired, you may not have the energy to do normal daily activities.	None	You are not weak or tired and you are able to do your usual physical, work, or social activities.
		Mild-to-moderate	You have trouble with strenuous physical activities such as exercising, climbing several flights of stairs, or running; you have trouble with moderate physical activities such as walking, housework, and shopping; you have some trouble doing your normal work activities; and you can take part in some normal social activities.
		Severe	You cannot work at a job and you need help taking care of yourself, including dressing, bathing, and cooking.
Stomach problems	Some cancer medicines can cause stomach problems which include diarrhea (loose or watery stool), nausea, and vomiting. Mild-to-moderate and severe stomach problems may limit your ability to do normal activities.	None	Not applicable
		Mild-to-moderate	Less than 7 stools per day; cramping; and nausea and vomiting. Managed with over-the-counter medicine and may need intravenous (IV) fluids.
		Severe	7 or more stools per day; cramping; nausea and vomiting; fever; extreme fluid loss (dehydration); bladder or bowel control problems; and can be life-threatening. May need IV fluids and your doctor may stop your RCC treatment or lower the dosage of your RCC medicine. If you have to stop taking the medicine, your cancer will get worse.
Sores in your mouth or throat (mucositis or stomatitis)	Some people who take cancer medicines develop sores in the mouth and throat. These sores can be mild-to-moderate, or severe.	None	Not applicable
		Mild-to-moderate	You experience soreness in the mouth or throat that (a) does not interfere with your ability to eat or drink or (b) makes it slightly hard to breathe or painful to eat solid foods. Managed with mild toothpaste; non-alcoholic mouthwash or salt water (or baking soda) rinse; over over-the-counter pain relievers.
		Severe	You experience soreness in the mouth or throat that makes it difficult to eat and drink; and can be life-threatening. Managed with a strong pain killer (opioid) and your doctor may stop your RCC treatment or lower the dosage of your RCC medicine. If you have to stop taking the medicine, your cancer will get worse.
Redness or sores on your hands and feet (hand–foot syndrome)	Some cancer medicines can cause the skin in the palms of your hands and the soles of your feet to become tender or red. This is called hand–foot syndrome. Hand–foot syndrome can be mild-to-moderate or severe.	None	Not applicable.
		Mild-to-moderate	You have dryness on the palms of your hands and on the soles of your feet. You may have peeling, bleeding, or blistering on the palms of your hands and on the soles of your feet. You may have mild pain on the palms of your hands and on the soles of your feet. You can do your normal activities. Managed with over-the-counter creams on hands and feet; ice packs for 15–20 min at a time; elevated hands and feet; over-the-counter pain relievers; antibiotic creams on open sores; and prescription creams on hands and feet.
		Severe	Your doctor may stop your RCC treatment or lower the dosage of your RCC medicine. If you have to stop taking the medicine, your cancer will get worse.
Chance of lung damage (pneumonitis)	There is a chance that cancer medicines could cause swelling and pain in the lungs leading to lung damage. The most common symptom is difficulty breathing, along with a cough and fever. However, it cannot be spread to another person. In most cases, this type of lung damage can be treated and cured.	No chance	Not applicable.
		5 out of 1000 (0.5%)
		10 out of 1000 (1.0%)
		20 out of 1000 (2.0%)
Chance of liver failure (hepatic impairment)	There is a chance that cancer medicines could cause liver failure. Liver failure occurs when large parts of the liver become damaged and cannot be repaired. A person who experiences liver failure can no longer work.	No chance	Not applicable.
		5 out of 1000 (0.5%)
		10 out of 1000 (1.0%)
		20 out of 1000 (2.0%)
How you take the medicine (mode of administration)	Cancer medicines can be taken as pills or infusions. Pills are taken by the mouth once or twice a day at home. With some medicines, you cannot eat before taking the pills. For other medicines, you can take the pills whether or not you have eaten. If you have an infusion, you will go to a doctor’s office, hospital, or clinic. A nurse will insert a needle into your arm that is attached to a bag filled with liquid medicine. It will take 30–60 min for all of the medicine to drip from the bag into your vein.	1 pill once a day with or without food 2 pills twice a day without food Infusion once a week.	Not applicable.

Figure 1.  Example of a choice question. Patients could click on the label to see the description for any medication feature. Prior to answering the choice questions, patients were asked to assume that (1) their RCC was metastatic, (2) they were starting a new medication, (3) this was the last medication they could take, and (4) all of their medical bills, including the cost of medications, were covered by health insurance. RCC = renal cell carcinoma.

Table 2.   Demographics and baseline characteristics of the sample.

Characteristic	Patients (n = 272)
Sex: n (%)
Male	127 (46.9)
Female	144 (53.1)
Mean age: years (SD)	57 (10.2)
Marital status: n (%)
Married	198 (73.6)
Other	71 (26.4)
Race: n (%)
White	251 (92.3)
Other	21 (7.7)
Highest education: n (%)
No college degree	92 (34.0)
Associate’s degree or higher	179 (66.1)
Employment status: n (%)
Employed	151 (55.7)
Unemployed	42 (15.5)
Other	78 (28.8)
Time since diagnosis: n (%)
Less than 2 years ago	95 (34.9)
At least 2 years ago	177 (65.1)
Had kidney surgery to treat RCC: n (%)
Yes	266 (97.8)
No	6 (2.2)
Has metastatic disease: n (%)
Yes	118 (43.4)
No/don’t know	154 (56.6)
Currently taking prescription medication: n (%)
Yes	73 (26.8)
No	199 (73.2)

RCC, renal cell carcinoma; SD, standard deviation.

Figure 2.  Preference weight graph. Preference weights for the eight attributes. Only relative differences matter when interpreting preference weights. The differences between adjacent weights indicate the relative importance of moving from one level of an attribute to an adjacent level of that attribute. The vertical distance between the best and worst level (that is, difference in the model coefficients) of each attribute is a measure of the overall mean relative importance of that attribute (over the ranges presented in the survey) to patients. The vertical bars surrounding each mean preference weight denote the 95% confidence intervals about the point estimate. If the confidence intervals do not overlap for adjacent levels in a particular attribute, the mean estimates are statistically different from each other at the 5% level of significance.

Figure 3.  Benefit equivalents. Benefit equivalents, as measured in months of progression-free survival, are measures of the relative importance of changes in toxicities that patients generally were willing to accept in exchange for accepting adverse events of varying severities. The horizontal bars surrounding each benefit equivalent denote the 95% confidence intervals about the point estimate. If the confidence intervals do not overlap for two benefit equivalents, the mean estimates are statistically different from each other at the 5% level of significance. The confidence intervals were estimated using the Krinsky-Robb procedure of 10,000 draws by dividing the difference in the preference weights of interest (e.g., an improvement in fatigue from ‘mild-to-moderate’ to ‘none’) by the relevant preference weight interval for progression-free survival.